1. Field of Invention
This invention relates to improvement in photoelectric generators for conversion of radiation into electrical energy and more specifically to such a semiconductor converter which will produce a relatively high power output under high intensity radiation levels and to methods of their manufacture.
2. Description of the Prior Art
There are a variety of energy sensitive devices that are extensively utilized to convert energy from one form to another; for example, the single crystal silicon photovoltaic cell, or silicon solar cell as it is herein referred to, has been successfully employed to convert incident solar radiation energy into electrical energy. Conventional silicon solar cells are well known and in extensive common use primarily because they provide relatively high conversion efficiencies compared to other energy conversion devices that are presently available. Because of this, the silicon solar cell has been used in the form of large flat solar arrays for terrestrial and space power aplications with electrical power output reaching the kilowatt level.
In conventional solar arrays, the cost of the silicon solar cells has been a predominant system cost factor and it has been recognized that cost savings can be achieved by using concentrators or reflectors to increase the intensity of incident radiation, thus providing increased power outputs per unit solar cell area. However, at high intensities, the conventional silicon solar cell in general suffers from the disadvantages of low efficiency, low voltage and current densities, and low level of radiation intensity at which saturation current (and power) is attained, so that the requisite efficiency of energy conversion at high radiation intensities is not obtained.
The prime means of maximizing the performance of the conventional silicon solar cell for high radiation intensities is to minimize its effective series resistance. The design considerations for silicon solar cells of the conventional type for high intensities have been exemplified in a technical paper by Lewis et al., (Solar Cell Characteristics at High Solar Intensities and Temperatures by C. A. Lewis and J. P. Kirkpatrick, IEEE 8th Photovoltaic Specialist Conference Proceedings pp. 123-134, Seattle, Wash., Aug. 4, 5, 6, 1970). Briefly, the major considerations are reviewed here and are shown in FIG. 1 which is a representative example of the conventional N on P silicon solar cell configuration in accordance with previous technology. The total internal resistance consists of the sum of various elements, namely, sheet resistance of the upper N layer (for an N on P slar cell), the bulk resistance of the P region, the resistance between the metallic contact and the semiconducting material, and the resistance of the metallic contact. These elements can be represented as lumped resistances except for the N region sheet resistance component which must be considered as a distributed parameter due to the nonuniform current distribution between the current collection metallic contact grids of the upper layer. The current passes uniformly through all the other elements of the silicon solar cell.
Parameters that must be considered in the design of high intensity solar cells of the conventional type are: (a) Sheet resistance and diffusion depth: Sheet resistance has been shown to be the dominating factor in the cell internal impedance. Deep diffusion is required in order to minimize the sheet resistance in this region. However, a decrease in cell sensitivity to the short wavelength energy will occur with deep diffusion because of the photon absorption characteristics of silicon; so a compromise must be made. The conventional solar cell has a diffusion depth in the neighborhood of 0.3 to 0.5 microns and recent improved cells, such as the so called "violet cell", were made by reducing the thickness of the diffused layer to the neighborhood of 0.1 micron. Accordingly, in Lewis and Kirkpatrick article, for the conventional solar cell optimized for high intensity up to 25 "suns", the diffused junction depth was in the region of 0.6 to 0.8 microns; which was considered a reasonable trade off between its spectral response and series resistance. (b) Bulk resistance: This must be made low to minimize the effective series resistance. However, diffusion length, minority carrier lifetime and therefore the short circuit current, I.sub.SC, will increase as bulk resistivity is increased while open circuit voltage, V.sub.OC, decreases. At the present time, base resistivities of 0.5 to 10 ohm-cm are commonly used in the conventional solar cell, yielding open circuit voltages in the range of 0.55 to 0.6 volts. Recent technology improvements in conventional solar cells are based on use of lower resistivities of 0.01 ohm-cm to yield even higher open circuit voltages. However, with all these factors in mind and knowing that sheet resistance is a dominating factor, the use of material with a base resistivity of less than 1 ohm-cm is not warranted in Lewis and Kirkpatrick for fabrication of high intensity conventional solar cells. (c) Cell thickness: This should be kept small in order to keep the bulk resistance at a low value. Also, only those excess carriers that are photogenerated within a depth of about a diffusion length from the junction are collected. Therefore, a cell thickness that is much greater than a minority carrier diffusion length will only add excess series resistance without increasing the short circuit current. On the other hand, a lower limit on the cell thickness is placed by limitations of material handling, mounting, thermal dissipation, etc. (d) Grid contacts: In the conventional silicon solar cell for high intensity, the grid design becomes extremely important. It is desired to maximize the current collection efficiency by covering the entire upper surface, but this is not feasible since the grid, being opaque, also blocks illumination and lowers the light gathering efficiency of the cell. Increasing the member of grid lines on the active front surface will minimize the grid contact resistance. Obviously, the closer the grid spacing the shorter the path length that must be traversed by the charge carriers and so the collection efficiency will be enhanced. A limit is imposed however, due to the reduction in the active area directly exposed to the sunlight, so a comprise is necessary. Therefore, for a high intensity solar cell more grids and thicker metalization is used. Whereas the conventional 2 cm.times.2 cm solar cell has 5 to 7 grids covering only 5 percent of the upper surface, the cell described in Lewis and Kirkpatrick for high intensity has 11 grids on 1 cm.times.2 on solar cell, with 10 percent of the upper surface being covered.
Radiation incident upon a conventional silicon solar cell may, generally speaking, be considered as consisting of the three fractions: the fraction of the incident radiation reflected at the surface which can be controlled to a large degree by anti-reflection coatings, the fraction absorbed in the semiconductor bulk, and the fraction that is transmitted through the cell and is absorbed in the back side ohmic contact. It is desired to have the highest fraction of incident radiation absorbed in the semiconductor bulk, since radiation which is reflected at the surface is unusable and that which is absorbed in the back side contributes only to heat.
Radiation which enters either the N or P regions of the junction is attenuated in an exponential manner characterized by the absorption coefficient .alpha. of the material. In the upper surface that is highly doped the absorption coefficient is high; consequently, minority carriers generated quickly recombine because the life time in this region is extremely short. Thus, this region is essentially a dead layer with very little output response for the short wavelength radiation that is incident on the silicon solar cell.
In the bulk region the absorbed photons will generate electron-hole pairs and those that lie within a diffusion length of the depletion region will likely contribute to output current because they have a chance to reach the junction region. Those generated beyond this region will recombine before they can be separated by the potential barrier associated with the junction region and not contribute to output. In general the spectral response of a typical silicon solar cell reflects the situation discussed above in that the response is degraded in the short wavelength region, peaks at about 0.85 to 0.9 micron, and is again degraded at longer wavelengths up to 1.1 micron, which is the limit with the silicon band gap.
Approximately 25 percent of the sun's energy on the Earth's surface is composed of wavelengths longer than 1.1 micron to which the silicon cannot respond. Thus, this energy is absored in the opaque back contact of the conventional silicon solar cell. The absorbed energy not converted to useful electrical energy is dissipated within the conventional solar cell. The majority of energy is absorbed in the upper surface region, being of short wavelength. The energy absorbed beyond a diffusion length of the junction is again converted to thermal dissipation. Thus, most of the absorbed energy results in thermal dissipation within the solar cell itself. There is a temperature gradient through the solar cell from the upper surface to the heat sink (assuming that the cell is cooled from the back side) that depends on the thermal conductivity and rate of energy flow per unit area.
In consideration of the design limitations discussed above, and the trade offs involved among the design parameters, it is concluded that no significant high intensity gains can be expected from technology improvements on the conventional silicon solar cell.
The high intensity solar cell described in this invention utilizes an edge illuminated or vertical junction structure that has been employed as a photovoltaic cell. Such a structure is described by J. P. Wise in U.S. Pat. No. 3,690,953. The vertical junction hardened solar cell taught by J. P. Wise's patent is not suitable for high intensity applications and is comparatively very expensive to fabricate. The structure taught by John H. Gault in U.S. Pat. No. 3,422,527 as a method of manufacture of high voltage solar cell bears a resemblance to the present invention. However, this invention differs in several important aspects. According to Gault's invention, wafers formed to have photogenerating characteristics by any standard well known manner are employed. In this invention the starting wafers are not found to have suitable photogenerating characteristics by any standard well known manner but differ considerably in that the diffusion depths and bulk resistivities employed would make a rather useless conventional cell by all standard criteria. In addition, according to Gault's invention each wafer is nickel plated on both surfaces, sintered, and then replaced with nickel prior to being assembled into a stack with interposed soldering wafers such as pure tin. Thereafter, the assemblage is placed in a furnace to cause soldering of the nickel plated surfaces. One distinguishing feature of the present invention is that aluminum is employed to allow the silicon wafers, which provides not only better thermal characterstics for high intensity but can be done at lower temperatures than the Ni sintering employed by Gault to help preserve the bulk lifetimes that are essential for efficiency. However, no information or claims are presented in Gault's patent as to aspects concerning high intensity.
According to the present invention, a structure is provided for a photovoltaic cell which results in improved characteristics for conversion of high intensity radiation. The high intensity solar cell in accordance with the invention eliminates most of the problems and design limitations associated with the prior art. The invention also covers methods of manufacturing series-connected arrays of high intensity solar cells at low cost, and a system of concentrating incident radiating into the most sensitive region of the high intensity solar cell for maximizing efficiency. The high intensity solar cell has demonstrated high efficiency and thermal stability under intense solar concentration levels in excess of 250 AMO sun levels.